UNBIASED EQUATION-ERROR APPROACH FOR EFFICIENT IIR SYSTEM IDENTIFICATION

摘要

Based on the equation-error approach, a weighted least squares algorithm with a generalized unit-norm constraint is developed for unbiased infinite impulse response (IIR) system identification in the presence of white input and/or output noise. Through using a weighting matrix, the proposed estimator can be considered as a generalization of the Koopmans-Levin method and it produces better estimation accuracy. Computer simulations are included to evaluate the estimator performance for IIR parameters in different conditions.